Our project

The experiment's goal is to study the behaviour of the buyers depending on the product's label. Our first phase is to analyze the given dataset and to evaluate which analysis to do. 325 partecipants joined in this experiment where we can see the distribution by age and salary with the presented histograms on the left. The buying simulation has been done on the pc, which is divided in two phases:

• Phase 1: the partecipants look at the catalog and select the products to buy without any nutritional label
• Phase 2: the partecipants look at the catalog and select the products to buy with any nutritional label

The scatterplot represents all the products in the catalog of the supermarcket based on the price by kg/l and their score FSA (Food Standards Agency). By pointing the relative product, we can see the informations about the product on the right side. We took in consideration four types of labels:

• Nutri-score: it's based on five colors that divides products in categories based on the score relative to the product's nutritional quality on 100g(energy, fat, saturated fat, sugar, salt, fiber)
• Traffic lights: red means the product is high in a nutrient and you should try to cut down, eat less often or eat smaller amounts. Amber means medium. If a food contains mostly amber, you can eat it most of the time. Green means low.
• RI-GDA: reference intakes (RIs) are a means of communicating maximum recommended nutrient intake to the public. Reference Intakes replaced the term Guideline Daily Amount (GDA), although the principles behind both are the same. The major difference is that GDAs existed for men, women and children; there is only one set of RIs for an average adult.
• Neutral: this label is like the RI-GDA but without the percentage of daily intake.

To evaluate which is the most effective label, we have done a deep analysis.
The bar chart on the right represents substantially the effectiveness of the label, measured with weighted parameter with the following formula:
Δdarkgreen*2 + Δlightgreen*1 + Δyellow*0 - Δorange*1 - Δred*2
The parameter "Δdarkgreen" is the result of the difference between the quantity of products' classified darkgreen bought in phase 2 and the quantity of products' classified darkgreen bought in phase 1. We have taken as reference the nutriscore 5c's colors to do a comparison having the same classification of the product. This result is represented by the charts underneath, where it is possible to filter by the type of label, by age and by salary.
From these analysis we can conclude that the label nutriscore 5c stimulates more buying products that are healthy. If we look at the neutral label (which is the label with less indications), it is noticeable that it doesn't boost sales for healthy products.